Re: [PATCH 19/54 v2] ASoC: switch over to use snd_soc_register_component() on mpc5200 i2s

[Stephen Warren <swarren@wwwdotorg.org>]

On 03/21/2013 07:19 PM, Kuninori Morimoto wrote:
> 
> Hi Stephen
> 
> Thank you for pointing it
> 
>>> diff --git a/sound/soc/fsl/mpc5200_psc_i2s.c b/sound/soc/fsl/mpc5200_psc_i2s.c
>>
>>> @@ -163,7 +167,8 @@ static int psc_i2s_of_probe(struct platform_device *op)
>>>  	if (rc != 0)
>>>  		return rc;
>>>  
>>> -	rc = snd_soc_register_dais(&op->dev, psc_i2s_dai, ARRAY_SIZE(psc_i2s_dai));
>>> +	rc = snd_soc_register_component(&op->dev, &psc_i2s_component,
>>> +					psc_i2s_dai, ARRAY_SIZE(psc_i2s_dai));
>>
>> ARRAY_SIZE(psc_i2s_dai)==1, so this switches from
>> snd_soc_register_dais() to snd_soc_register_dai(). Will that be a problem?
> 
> I didn't notice about it.
> 
> arch/powerpc/boot/dts/lite5200.dts seems the user of this driver,
> but, it was comment-outed.
> And it is using DT, not name matching ?

This one required some more in-depth investigation.

Certainly, there don't appear to be any instance of the device-name or
DAI name string mpc5200-psc-i2s anywhere in the kernel.

However, I do see:

> ./mpc8610_hpcd.c:213:	machine_data->dai[0].cpu_dai_name = dev_name(&ssi_pdev->dev);
> ./p1022_ds.c:223:	mdata->dai[0].cpu_dai_name = dev_name(&ssi_pdev->dev);
> ./p1022_rdk.c:226:	mdata->dai[0].cpu_dai_name = dev_name(&ssi_pdev->dev);
> ./imx-sgtl5000.c:151:	data->dai.cpu_dai_name = dev_name(&ssi_pdev->dev);

That's a bit icky; DT-based systems should use .cpu_of_node, not
.cpu_dai_name...

Anyway, as far as I can tell from any of the in-kernel device trees,
this particular mpc5200-psc-i2s device isn't actually instantiated
anywhere. I guess the only users are removed or out-of-tree.

Equally, of those 4 files, the compatible value they're instantiated
from doesn't exist in any in-kernel DT files, except for imx-sgtl5000.c,
and in that case the CPU phandle points at something the definitely
isn't the MPC5200 PSC device.

So, I guess this if probably fine...